Example Siamese Network with CNN with Tensorflow 2 and Keras in Python 3
€30-250 EUR
Paid on delivery
I need an example of an Siamese Network with CNN with certain guidelines.
Prerequisites: The code should work with tensorflow 2.3 and 2.4 on a linux and Windows machine with tensorflow GPU support
1) Source images are organized in folders as typically used for classification, e.g.
test/spoon
test/fork
test/knive
train/spoon
train/fork
train/knive
2) Equal (e.g fork/fork) and unequal (e.g fork/spoon) pairs have to be generated for training and test. Parameters controlling the amount of used pairs should be in the code
3) Base network
- The parallel network should be exchangeable easily. A funtion generating the base network should be in the code (e.g. create_base_network(input_shape)).
- It should be possible to use models from [login to view URL] as base networks easily
- The network should work with different image sizes (parameter image size..., of course, must fit to the models above)
- Image preprocessing should be done via Keras preprocess_input, e.g with tf.keras.applications.resnet.preprocess_input for ResNet50
- Data augmentation should be possible via the network, e.g.
data_augmentation = [login to view URL]([[login to view URL]('horizontal_and_vertical'), [login to view URL](0.02, 0.02)]
x = data_augmentation(input)
x = preprocess_input(x)
...
It should be possible to switch data augmentation on and off (e.g. on for train data, off for test data)
- In the example please use ResNet50 with data augmentation
4) Head network
- Should be exchangable easily (function create_head_model(embedding_shape))
5) Training
- Callback functions should be used
- Models should be saved
- Except normal output during training, it should be possible to view the training via tensorboard
6) Use of model
- A trained model should be used with the input of two images as an example of usage of the model
Hint: Maybe something like [login to view URL] could be an idea of a template.
Here the link for the images for the sample data:
[login to view URL]
Project ID: #29616817
About the project
Awarded to:
Hi - I am a forbes featured AI developer. I've worked with big companies with solutions comprising of Tensorflow. I can help in finishing your task within two days (Over the weekend). Let me if I can be of any help. Ch More
12 freelancers are bidding on average €224 for this job
Hi, Greetings! ✅ I have checked and gone through the project details and I am very skilled in it. ✅Completed Time: In project deadline We have worked on 640 + Projects. I have 6 + years of the experience in sam More
Ping me- I am clear with the task. I have done similar project before. Hi i am an Senior Python , Tensorflow , Deep learning , keras Expert. As 9+ years experiences in these field. I can give good quality work. I have More
Hello! I've many experiences in CNN models by tensorflow or keras. ResNet50 is known to be the good model for classification of images. So Inception and Xception are. Siamese Network is used to match or verify 2 object More
Hello, I am very familiarized with the requirements of your projects. And it can be done really fast. Let's connect over chat to discuss more on this. Thanks,
Hi, Greetings for the day! We are a team of passionate Machine Learning Engineers based in Mumbai, India, who have developed and deployed complex AI solutions using Machine Learning and Computer Vision. We have a g More
***Example Siamese Network with CNN with Tensorflow 2 and Keras in Python 3*** Hi there, I am the best match for this project as I have strong experience with Python, Machine Learning (ML), Artificial Intelligence, T More